Re: mmotm 2021-05-12-21-46 uploaded (arch/x86/mm/pgtable.c)

[Randy Dunlap]

On 5/13/21 1:47 PM, Andrew Morton wrote:
> On Thu, 13 May 2021 19:09:23 +0200 Christophe Leroy <christophe.leroy@xxxxxxxxxx> wrote:
> 
>>
>>
>>> on i386:
>>>
>>> ../arch/x86/mm/pgtable.c:703:5: error: redefinition of ‘pud_set_huge’
>>>   int pud_set_huge(pud_t *pud, phys_addr_t addr, pgprot_t prot)
>>>       ^~~~~~~~~~~~
>>> In file included from ../include/linux/mm.h:33:0,
>>>                   from ../arch/x86/mm/pgtable.c:2:
>>> ../include/linux/pgtable.h:1387:19: note: previous definition of ‘pud_set_huge’ was here
>>>   static inline int pud_set_huge(pud_t *pud, phys_addr_t addr, pgprot_t prot)
>>>                     ^~~~~~~~~~~~
>>> ../arch/x86/mm/pgtable.c:758:5: error: redefinition of ‘pud_clear_huge’
>>>   int pud_clear_huge(pud_t *pud)
>>>       ^~~~~~~~~~~~~~
>>> In file included from ../include/linux/mm.h:33:0,
>>>                   from ../arch/x86/mm/pgtable.c:2:
>>> ../include/linux/pgtable.h:1391:19: note: previous definition of ‘pud_clear_huge’ was here
>>>   static inline int pud_clear_huge(pud_t *pud)
>>>                     ^~~~~~~~~~~~~~
>>
>> Hum ...
>>
>> Comes from my patch 
>> https://patchwork.ozlabs.org/project/linuxppc-dev/patch/5ac5976419350e8e048d463a64cae449eb3ba4b0.1620795204.git.christophe.leroy@xxxxxxxxxx/
>>
>> But, that happens only if x86 defines __PAGETABLE_PUD_FOLDED. And if PUD is folded, then I can't 
>> understand my it has pud_set_huge() and pud_clear_huge() functions.
> 
> Probably because someone messed something up ;)
> 
> Let's try this.
> 
> --- a/arch/x86/mm/pgtable.c~mm-pgtable-add-stubs-for-pmd-pub_set-clear_huge-fix
> +++ a/arch/x86/mm/pgtable.c

That also works_for_me.

Acked-by: Randy Dunlap <rdunlap@xxxxxxxxxxxxx> # build-tested

thanks.
-- 
~Randy